Ultrasound Obstet Gynecol 2015; 46: 616–622 Published online 5 October 2015 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/uog.14825

Accuracy of three-dimensional ultrasound compared ¨ with magnetic resonance imaging in diagnosis of Mullerian duct anomalies using ESHRE–ESGE consensus on the classification of congenital anomalies of the female genital tract ´ B. GRAUPERA, M. A. PASCUAL, L. HERETER, J. L. BROWNE, B. UBEDA, I. RODRI´GUEZ and C. PEDRERO Department of Obstetrics, Gynecology and Reproduction, Hospital Universitari Quiron Dexeus, Barcelona, Spain

¨ K E Y W O R D S: 3D-US; congenital uterine anomalies; magnetic resonance imaging; Mullerian duct anomalies; three-dimensional ultrasonography; uterine abnormalities; uterine malformations

ABSTRACT Objective To establish the accuracy of three-dimensional ultrasound (3D-US), compared with magnetic resonance imaging (MRI), for diagnosing uterine anomalies, using the European Society of Human Reproduction and Embryology–European Society for Gynaecological Endoscopy (ESHRE–ESGE) consensus on the classification of congenital anomalies of the female genital tract. Methods Sixty women with uterine anomalies suspected after examination by conventional two-dimensional ultrasound were evaluated with 3D-US and MRI. These data were analyzed retrospectively to confirm the presence and type of uterine malformation in accordance with the ESHRE–ESGE consensus. Sensitivity, specificity and positive (PPV) and negative (NPV) predictive values were calculated, using MRI as the gold standard, and agreement between the two methods was evaluated by kappa index. Results Compared with MRI, for the diagnosis of normal uteri, 3D-US had a sensitivity of 83.3%, specificity of 100%, PPV of 100%, NPV of 98.2% and kappa index of 0.900. For dysmorphic uteri and for hemi-uteri, the sensitivity, specificity, PPV and NPV were all 100%, and kappa was 1.00. For septate uteri, the sensitivity was 100%, specificity was 88.9%, PPV was 95.5%, NPV was 100% and kappa was 0.918. For bicorporeal uteri, the sensitivity was 83.3%, specificity was 100%, PPV was 100%, NPV was 98.2% and kappa was 0.900. Conclusions 3D-US is highly accurate for diagnosing uterine malformations, having a good level of agreement

with MRI in the classification of different anomaly types based on the ESHRE–ESGE consensus. Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

INTRODUCTION ¨ Congenital uterine anomalies, also known as Mullerian duct anomalies (MDA) due to their embryological origin ¨ in the Mullerian ducts, result from isolated or combined alterations involved in embryogenic development in the uterus. The prevalence of these malformations tends to vary depending on the population studied, occurring in around 5.5% of the general population and in up to around 24.5% of infertile patients with a history of miscarriage1 . Until now, the system most commonly used for classification of uterine anomalies has been that of the American Fertility Society (AFS), published in 19882 . Recently, however, a new consensus was established for the classification of congenital malformations of the female genital tract3 . Several different methods are used for diagnosing uterine anomalies. The first is conventional two-dimensional transvaginal ultrasound (2D-US)4 . Hysteroscopy is useful for evaluating the uterine cavity, while laparoscopy is useful in assessment of the external uterine contour; hysteroscopy and laparoscopy have been considered to be the gold standard in the diagnosis of MDA5 – 7 . Magnetic resonance imaging (MRI) and threedimensional ultrasound (3D-US) hold a distinct advantage over other techniques, in that they provide simultaneously information about both the external contour and

Correspondence to: Dr B. Graupera, Department of Obstetrics, Gynecology and Reproduction, Hospital Universitari Quiron Dexeus, Gran Via Carles III, 71–75, 08028 Barcelona, Spain (e-mail: [email protected]) Accepted: 13 February 2015

Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

ORIGINAL PAPER

3D-US to diagnose MDA the cavity of the uterus. MRI has proved highly accurate in the diagnosis of MDA and has been considered the gold standard imaging technique8 – 11 . However, 3D-US is a non-invasive, reproducible and low-cost technique that has the added advantage of being able to capture the third, or coronal, plane. This plane, obtained using volumetric reconstruction, provides an exceptionally clear frontal view of the uterus and its anatomical details12 . Accordingly, 3D-US could be a viable alternative to MRI13 . There have been previous studies comparing 3D-US with MRI in the diagnosis of uterine malformations using the AFS classification14 – 16 . The objective of this study was to establish the accuracy of 3D-US, compared with the gold standard of MRI, for diagnosing uterine anomalies, using the European Society of Human Reproduction and Embryology–European Society for Gynaecological Endoscopy (ESHRE–ESGE) consensus on the classification of congenital anomalies of the female genital tract3 .

METHODS This was a retrospective analysis of data collected prospectively from a consecutive series of premenopausal women attending our institution between September 2004 and July 2009. Sixty patients with uterine malformation suspected on 2D-US were examined with 3D-US and MRI. These results have been analyzed previously as part of a doctoral thesis17 , with uterine anomalies classified according to the AFS2 and both diagnostic techniques being compared. In the present study, after obtaining institutional review board approval, we analyzed the same data retrospectively according to the ESHRE–ESGE consensus3 , with the observer being blinded to the results of each technique while analyzing the results of the other.

Three-dimensional ultrasound All ultrasound examinations were performed using a GE Healthcare Voluson 730 Expert (Milwaukee, WI, USA) ultrasound machine, equipped with a variable-frequency (2.9–10 MHz) transvaginal probe. The settings used during the exam were constant for all patients: power, 100%; gain, 6. The volumes obtained were analyzed offline at a personal computer workstation with 4DView (version 5.0) software developed by GE Medical Systems Kretztechnik GMBH & Co OHG (Zipf, Austria). The uterine volume was obtained from the sagittal plane showing the uterus from the cervix to the fundus, with the endometrial line in a horizontal direction and perpendicular to the ultrasonic beam. In any uterus with an increased transverse diameter, an additional acquisition was made from a transverse section. Automatic volume acquisition was performed with a machine setting of normal quality using a 90◦ sweep angle to obtain a multiplanar view of the uterus. On completing the procedure, we checked that the whole uterus had been captured in all three planes. If not, volume acquisition was repeated and, if necessary, the volume area was modified.

Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

617

The volume data were observed in three orthogonal planes in addition to the reconstruction. Rotations and translations were carried out on each axis of each plane passing through the center point at which the three perpendicular planes intersected. In general, surface and transparent modes were used to provide a clear view of the uterus and uterine cavity. The total time required to diagnose uterine malformations using 3D-US included the time for volume data acquisition (< 1 min) plus the time for post-processing (usually 3 min). In some cases the time could reach 10 min, depending on the complexity of the abnormality.

Magnetic resonance imaging MRI examinations were performed with a GE Signa 1-T (Milwaukee, WI, USA) scanner from the start of the study until July 2007 and with a 1.5-T scanner with a phased-array body surface coil from July 2007 onwards. The protocol for performing MRI in women with suspected uterine abnormalities included the following sequences: axial T1-weighted fast spin-echo (FSE), axial FSE T1-weighted with fat saturation, axial T2-weighted FSE, sagittal T2-weighted FSE, oriented according to the uterine axis, and coronal T2-weighted FSE, oriented parallel to the major uterine axis. It generally took about 15 min to perform MRI for suspected uterine anomalies, plus 15 min for post-processing.

Evaluation of findings We evaluated retrospectively 3D-US volumes and MRI data to confirm the presence of uterine malformations and classified the different types of anomalies in accordance with the ESHRE–ESGE consensus3 . The 3D-US analysis was performed by three expert examiners (B.G., M.A.P., L.H.) and MRI data were evaluated by two expert examiners (J.L.B., B.U.). The 3D-US results were then compared with those of MRI. At the time of evaluation of the 3D-US volumes, neither the conventional 2D-US findings nor the MRI results were known. Analysis of uterine morphology was performed on a standardized reconstructed plane, displaying the coronal plane of the uterus with the interstitial portion of both Fallopian tubes as points of reference18 . According to the ESHRE–ESGE consensus, all cases of normal uterus were classified as U0. A normal uterus was defined as any uterus having an interostial line with an internal indentation at the fundal midline not exceeding 50% of the uterine wall thickness. Class U1, or dysmorphic uterus, included all cases with a normal uterine outline but with an abnormal shape of the uterine cavity excluding septa. Class U2, or septate uterus, was defined as a uterus with normal outline and an internal indentation at the fundal midline exceeding 50% of the uterine wall thickness (Figure 1); a septum may partially (Class U2a) or completely (Class U2b) divide the uterine cavity. Class U3, or bicorporeal uterus, was defined as a uterus with an abnormal fundal outline along with

Ultrasound Obstet Gynecol 2015; 46: 616–622.

Graupera et al.

618

Figure 1 Three-dimensional ultrasound in the diagnosis of uterine anomaly Class U2a: partial septate uterus. (a) Normal outline morphology with septum dividing uterine cavity partially. (b) Uterine wall thickness measured from external contour to interostial tube line (caliper 2, 16 mm). (c) Cavity is divided partially by septum (caliper 2, 26 mm) at fundal midline, internal indentation being >50% of the uterine wall thickness.

Figure 2 Three-dimensional ultrasound in the diagnosis of uterine anomaly Class U3: bicorporeal uterus. (a) Class U3a: partial bicorporeal uterus. Abnormal uterine outline with external indentation >50% of uterine wall thickness. The indentation is measured from the line drawn between both horns to the deepest point of the fundal indentation (caliper 2). In this example this distance is 11 mm, being >50% of the uterine wall thickness. (b) Class U3b: complete bicorporeal uterus. Abnormal uterine outline with external indentation completely dividing uterine corpus up to the level of the cervix. (c) Class U3c: bicorporeal septate uterus. Abnormal uterine outline with external indentation > 50% of uterine wall thickness. The thickness of the septum at the fundal midline is >150% of the uterine wall thickness, with septum dividing uterine cavity completely.

the presence of an external indentation at the fundal midline exceeding 50% of the uterine wall thickness; this indentation could divide the uterine cavity partially (Class U3a) or completely (Class U3b). This class also included bicorporeal septate uterus (Class U3c), in which the width of the midline fundal indentation exceeded the uterine wall thickness by 50% (Figure 2). Class U4, or hemi-uterus, was defined as a unilateral uterine development with (Class U4a) or without (Class U4b) rudimentary cavity. Class U5 incorporated all cases of uterine aplasia. Class U6 was reserved for cases that remained unclassified3 .

Statistical analysis For each class of uterine malformation, we analyzed the accuracy of 3D-US using findings on MRI as the gold standard. The relationship between the variables was evaluated using Pearson’s chi-square or Fisher’s exact test. The diagnostic capability of 3D-US was evaluated using sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). Agreement between the different methods was evaluated using the kappa index. All tests were bilateral, with significance level of α = 0.05. Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

RESULTS Between September 2004 and July 2009, 60 women (mean (SD) age, 31.7 (4.7) years) with uterine malformation suspected at 2D-US were evaluated with 3D-US and MRI. MRI was performed within 3 months following 3D-US examination. According to 3D-US: five (8.3%) patients were Class U0, normal uterus; one (1.7%) patient was Class U1, dysmorphic uterus; 44 (73.3%) patients were Class U2, septate uterus; five (8.3%) patients were U3, bicorporeal uterus; and five (8.3%) patients were U4, hemi-uterus (Figures 3–5). According to MRI, six (10%) patients were Class U0, normal uterus; one (1.7%) patient was Class U1, dysmorphic uterus; 42 (70%) patients were Class U2, septate uterus; six (10%) patients were Class U3, bicorporeal uterus; and five (8.3%) patients were Class U4, hemi-uterus (Figures 6 and 7). Table 1 shows the accuracy of 3D-US in the diagnosis of the different types of uterine anomaly. For the diagnosis of Class U2, septate uterus, the most common anomaly, the sensitivity was 100%, specificity was 88.9% (95% CI, 74.4–100), PPV was 95.5% (95% CI, 89.3–100) and NPV was 100%. Concordance between 3D-US and MRI showed a kappa value of 0.918 (P < 0.001) in the diagnosis of septate uterus.

Ultrasound Obstet Gynecol 2015; 46: 616–622.

3D-US to diagnose MDA

619

Figure 3 Three-dimensional ultrasound diagnosis according to ESHRE–ESGE consensus: (a) Class U0, normal uterus; (b) Class U1, dysmorphic uterus; (c) Class U2a, partial septate uterus; (d) Class U2b, complete septate uterus; (e) Class U3a, partial bicorporeal uterus; (f) Class U3b, complete bicorporeal uterus; (g) Class U4, hemi-uterus. Table 1 Accuracy of three-dimensional ultrasound compared with magnetic resonance imaging in the diagnosis of different types of uterine anomaly according to the European Society of Human Reproduction and Embryology–European Society for Gynaecological Endoscopy consensus Uterine anomaly Normal (Class U0) Dysmorphic (Class U1) Septate (Class U2) Bicorporeal (Class U3) Hemi-uterus (Class U4)

Sensitivity

Specificity

PPV

NPV

κ

83.3 (53.5–100) (5/6) 100 (1/1) 100 (42/42) 83.3 (53.5–100) (5/6) 100 (5/5)

100 (54/54) 100 (59/59) 88.9 (74.4–100) (16/18) 100 (54/54) 100 (55/55)

100 (5/5) 100 (1/1) 95.5 (89.3–100) (42/44) 100 (5/5) 100 (5/5)

98.2 (94.6–100) (54/55) 100 (59/59) 100 (16/16) 98.2 (94.6–100) (54/55) 100 (55/55)

0.900* 1.000* 0.918* 0.900* 1.000*

Data are given as % (95% CI) (n/n) or % (n/n). *P < 0.001. κ, kappa index; NPV, negative predictive value; PPV, positive predictive value.

There were only two discrepancies between 3D-US and MRI; in both cases, 3D-US diagnosed septate uterus (Class U2), while MRI diagnosed a normal uterus (Class U0) in one and a bicorporeal uterus (Class U3) in the other.

DISCUSSION 3D-US is a valuable method for the diagnosis of uterine anomalies, not least because of its capacity to demonstrate the coronal plane. Like MRI, it reveals information on both the morphology of the uterine cavity and its relationship with the external uterine contour. Obtaining an accurate diagnosis of uterine anomalies is important due to their association with adverse reproductive outcome7,19 – 21 . 3D-US, unlike 2D-US, which only provides indirect findings regarding the presence of uterine malformations, allows differential diagnosis between the different types of uterine anomaly. Additionally, 3D-US allows quantification of the severity of the anomaly and provides data that are fundamental to understanding a patient’s reproductive prognosis and to

Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

proper selection of an individualized therapeutic strategy. Finally, it has also been demonstrated to have good intra- and interobserver reproducibility in the diagnosis of uterine anomalies22 . The accuracy of 3D-US in diagnosing uterine malformations has been the subject of many studies since the first was published in 19954 . These studies initially compared 3D-US with hysterosalpingography, hysteroscopy and laparoscopy23 – 28 and found it was very good in terms of its diagnostic capabilities. Deutch et al.29 compared MRI and 3D-US results with surgical results, finding that 3D-US had a sensitivity of 100% in both diagnosis and categorization of uterine abnormalities. Caliskan et al.15 compared the diagnostic accuracy of 3D-US with that of laparoscopy and hysteroscopy, and also with MRI, in both the first and second phases of the menstrual cycle, recording a sensitivity of 94.7% for 3D-US in the follicular phase and 100% in the luteal phase and a specificity of 75% in the follicular phase and 93.7% in the luteal phase; no arcuate uteri (Type VII, AFS classification) were diagnosed. Bermejo et al.14

Ultrasound Obstet Gynecol 2015; 46: 616–622.

620

Figure 4 Three-dimensional ultrasound image of complete septate uterus (Class U2b). Note septum fully dividing uterine cavity up to level of internal cervical os.

Figure 5 Three-dimensional ultrasound image of bicorporeal septate uterus (Class U3c). Note external fundal indentation (arrow) dividing uterine corpus above level of cervix.

found agreement between 3D-US and MRI in diagnosing uterine anomalies, with a kappa index of 0.880. Faivre et al.16 demonstrated accuracy of 100% for 3D-US in the differential diagnosis between septate and bicornuate uterus, compared with MRI. 3D-US has proven to be just as effective as MRI in the diagnosis of uterine malformations. It is also more economical and better tolerated by patients13 . Precise diagnostic imaging should decrease the need for invasive

Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

Graupera et al.

Figure 6 Magnetic resonance image of complete septate uterus (Class U2b), demonstrating septum completely dividing uterine cavity.

Figure 7 Magnetic resonance image of bicorporeal septate uterus (Class U3c), showing external fundal indentation (arrow) partially dividing uterine corpus above level of cervix.

techniques such as laparoscopy and hysteroscopy, which could potentially be reserved for those women requiring therapeutic intervention30 . 3D-US should be used when there is suspicion of MDA, MRI should be used in doubtful or complex cases and surgery reserved for malformations that will benefit from this treatment14 . To our knowledge, this is the first study comparing 3D-US with MRI for analysis of uterine anomalies using the ESHRE–ESGE consensus. Previous studies have proven 3D-US to be an effective technique, with good results for sensitivity and specificity in both diagnosing

Ultrasound Obstet Gynecol 2015; 46: 616–622.

3D-US to diagnose MDA

621

¨ Figure 8 Suspected Mullerian duct anomaly with discrepancy of diagnosis: partial septate uterus on three-dimensional ultrasound (3D-US) (a) and normal uterus on magnetic resonance imaging (MRI) (b). (a) 3D-US showing uterine cavity partially divided by septum at fundal midline, exceeding 50% of uterine wall thickness. Length of septum, 8 mm (caliper 2). (b) At MRI, this fundal indentation was not considered to be sufficient to diagnose septate uterus, and normal morphology was reported. Arrows indicate tubal ostia.

¨ Figure 9 Suspected Mullerian duct anomaly with discrepancy of diagnosis: septate uterus on three-dimensional ultrasound (3D-US) (a) and bicorporeal uterus on magnetic resonance imaging (MRI) (b). (a) 3D-US showing minimal indentation in external fundal contour (3 mm, caliper 2) < 50% of myometrial thickness (8 mm, caliper 3), consistent with diagnosis of septate uterus. (b) MRI reported that indentation (arrow) was consistent with bicorporeal uterus.

and classifying malformations. This study included ESHRE–ESGE Class U0, U1, U2, U3 and U4 uterine anomalies, finding a good level of concordance between 3D-US and MRI. The ESHRE–ESGE consensus provides objective parameters that allow for precise categorization of uterine abnormalities into one of six classes (U0–U5). It also includes a seventh class (U6) for unclassified cases, infrequent abnormalities, subtle changes and combined pathologies that can not be included in Classes U0–U5.

Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

We found only two discrepancies between 3D-US and MRI in the categorization of uterine anomalies. In both cases, these discrepancies were due to minor differences in the measurement of indentation at the fundal midline: the internal indentation in the first case and the external indentation in the second case. In the first case, 3D-US showed a normal outline and an internal indentation at the fundal midline exceeding 50% of the uterine wall thickness and partially affecting the cavity, leading to a diagnosis of partial septate uterus (Class U2a).

Ultrasound Obstet Gynecol 2015; 46: 616–622.

622

MRI, however, revealed normal morphology (Class U0) (Figure 8). Diagnostic hysteroscopy was not performed. In the second case, 3D-US showed minimal external indentation at the fundal midline of less than 50% of the uterine wall thickness, suggesting septate uterus (Class U2a), while MRI reported an indentation greater than 50% of the myometrial wall thickness, with a diagnosis of bicorporeal uterus (Class U3) (Figure 9). Diagnostic hysteroscopy was performed in this case, confirming the ultrasound diagnosis of septate uterus, and the patient underwent surgical hysteroscopy to remove the septum. Congenital uterine anomalies result from isolated or combined alterations involved in embryogenic development in the uterus and can present characteristics intermediate between two types of malformation. We think that the ESHRE–ESGE consensus, providing objective parameters with which to classify uterine anomalies, allows 3D-US to distinguish accurately between these different types. Ludwin et al.31 encountered problems in the diagnosis of uterine congenital anomalies according to the ESHRE–ESGE compared with the AFS classification2 . According to Ludwin et al., septate uterus is diagnosed in cases in which the uterus fulfils previous morphometric criteria for arcuate or in some cases normal uteri. On the other hand, Grimbizis et al.32 asserted that the new system gives a unique opportunity to achieve an objective estimation of the clinical consequences related to the various degrees of uterine deformity. One of the strengths of our study is the number of patients with MRI correlation. Only one previous study has comparable data14 . Our study had some limitations. First, there was a sampling bias, since 3D-US is usually performed only on those patients whose 2D-US findings suggest the presence of a uterine malformation. Second, there was a limited number of some uterine anomaly types, which will have affected the validity of diagnostic accuracy data. Finally, this study was a retrospective analysis, albeit providing a new classification of data collected prospectively. In conclusion, according to our results, 3D-US is accurate in diagnosing uterine malformations, showing a good level of agreement with MRI in the classification of the different anomaly types when using the ESHRE–ESGE consensus.

ACKNOWLEDGMENTS This study was conducted under the auspices of the ` Catedra d’ Investigacio´ en Obstetr´ıcia i Ginecologia of ` the Universitat Autonoma de Barcelona. B.G. received a grant from Fundacio´ Dexeus Salut de la Dona.

REFERENCES 1. Chan YY, Jayaprakasan K, Zamorra J, Thornton JG, Raine-Fenning N, Coomarasamy A. The prevalence of congenital uterine anomalies in unselected and high-risk populations: a systematic review. Hum Reprod Update 2011; 17: 761–771.

Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

Graupera et al. 2. The American Fertility Society classifications of adnexal adhesions, distal tubal ¨ obstruction, tubal occlusion secondary to tubal ligation, tubal pregnancies, Mullerian anomalies and intrauterine adhesions. Fertil Steril 1988; 49: 944–955. 3. Grimbizis GF, Gordts S, Di Spiezio Sardo A, Brucker S, De Angelis C, Gergolet M, ¨ Li TC, Tanos V, Brolmann H, Gianaroli L, Campo R. The ESHRE–ESGE consensus on the classification of female genital tract congenital anomalies. Hum Reprod 2013; 28: 2032–2044. 4. Jurkovic D, Geipel A, Gruboeck K, Jauniaux E, Natucci M, Campbell S. Three-dimensional ultrasound for the assessment of uterine anatomy and detection of congenital anomalies: comparison with hysterosalpingography and two-dimensional sonography. Ultrasound Obstet Gynecol1995; 5: 233–237. 5. Randolph JF,Jr, Ying YK, Maier DB, Schmidt CL, Riddick DH. Comparison of real-time ultrasonography, hysterosalpingography, and laparoscopy/hysteroscopy in the evaluation of uterine abnormalities and tubal patency. Fertil Steril 1986; 46: 828–832. 6. Fedele L, Ferrazzi E, Dorta M, Vercellini P, Candiani GB. Ultrasonography in the differential diagnosis of ”double” uteri. Fertil Steril 1988; 50: 361–364. 7. Homer HA, Li TC, Cooke ID. The septate uterus: a review of management and reproductive outcome. Fertil Steril 2000; 73: 1–14. ¨ 8. Troiano RN. Magnetic resonance imaging of Mullerian duct anomalies of the uterus. Top Magn Reson Imaging 2003; 14: 269–279. 9. Wagner BJ, Woodward PJ. Magnetic resonance evaluation of congenital uterine anomalies. Semin Ultrasound CT MR 1994; 15: 4–17. 10. Pellerito JS, McCarthy SM, Doyle MB, Glickman MG, DeCherney AH. Diagnosis of uterine anomalies: relative accuracy of MR imaging, endovaginal sonography, and hysterosalpingography. Radiology 1992; 183: 795–800. ¨ 11. Doyle MB. Magnetic resonance imaging in Mullerian fusion defects. J Reprod Med 1992; 37: 33–38. ´ 12. Pascual MA, Hereter L, Graupera B, Fernandez-Cid M, Dexeus S. Ecograf´ıa 3D/4D en ginecolog´ıa: t´ecnica y metodolog´ıa. Prog Obstet Ginecol 2006; 49: 263–271. 13. Deutch TD, Abuhamad AZ. The role of 3-dimensional ultrasonography and magnetic ¨ resonance imaging in the diagnosis of Mullerian duct anomalies: a review of the literature. J Ultrasound Med 2008; 27: 413–423. ´ 14. Bermejo C, Mart´ınez Ten P, Cantarero R, Diaz D, P´erez Pedregosa J, Barron ´ E, Labrador E, Ruiz Lopez L. Three-dimensional ultrasound in the diagnosis ¨ of Mullerian duct anomalies and concordance with magnetic resonance imaging. Ultrasound Obstet Gynecol 2010; 35: 593–601. 15. Caliskan E, Ozkan S, Cakiroglu Y, Sarisoy HT, Corakci A, Ozeren S. Diagnostic ¨ accuracy of real-time 3D sonography in the diagnosis of congenital Mullerian anomalies in high-risk patients with respect to the phase of the menstrual cycle. J Clin Ultrasound 2010; 38: 123–127. 16. Faivre E, Fernandez H, Deffieux X, Gervaise A, Frydman R, Levaillant JM. Accuracy of three-dimensional ultrasonography in differential diagnosis of septate and bicornuate uterus compared with office hysteroscopy and pelvic magnetic resonance imaging. J Minim Invasive Gynecol 2012; 19: 101–106. 17. Graupera B. Validacion de las ´ de la ecograf´ıa 3D como t´ecnica diagnostica ´ ` malformaciones uterinas de origen mulleriano. Thesis. Universitat Autonoma de ¨ Barcelona, Barcelona, 2012. 18. Salim R, Regan L, Woelfer B, Backos M, Jurkovic D. A comparative study of the morphology of congenital uterine anomalies in women with and without a history of recurrent first trimester miscarriage. Hum Reprod 2003; 18: 162–166. 19. Tulandi T, Arronet GH, McInnes RA. Arcuate and bicornuate uterine anomalies and infertility. Fertil Steril 1980; 34: 362–364. 20. Acien P. Reproductive performance of women with uterine malformations. Hum Reprod 1993; 8: 122–126. 21. Raga F, Bauset C, Remohi J, Bonilla-Musoles F, Simon C, Pellicer A. Reproductive ¨ impact of congenital Mullerian anomalies. Hum Reprod 1997; 12: 2277–2281. 22. Salim R, Woelfer B, Backos M, Regan L, Jurkovic D. Reproducibility of three-dimensional ultrasound diagnosis of congenital uterine anomalies. Ultrasound Obstet Gynecol 2003; 21: 578–582. ¨ 23. Raga F, Bonilla-Musoles F, Blanes J, Osborne NG. Congenital Mullerian anomalies: diagnostic accuracy of three-dimensional ultrasound. Fertil Steril 1996; 65: 523–528. ¨ 24. Wu MH, Hsu CC, Huang KE. Detection of congenital Mullerian duct anomalies using three-dimensional ultrasound. J Clin Ultrasound 1997; 25: 487–492. 25. Kupesic S, Kurjak A. Septate uterus: detection and prediction of obstetrical complications by different forms of ultrasonography. J Ultrasound Med 1998; 17: 631–636. 26. Kupesi´c S, Kurjak A, Skenderovic S, Bjelos D. Screening for uterine abnormalities by three-dimensional ultrasound improves perinatal outcome. J Perinat Med 2002; 30: 9–17. 27. Momtaz MM, Ebrashy AN. Three-dimensional ultrasonography in the evaluation of the uterine cavity. Middle East Fertil Soc J 2007; 12: 41–46. 28. Ghi T, Casadio P, Kuleva M, Perrone AM, Savelli L, Giunchi S, Meriggiola MC, Gubbini G, Pilu G, Pelusi C, Pelusi G. Accuracy of three-dimensional ultrasound in diagnosis and classification of congenital uterine anomalies. Fertil Steril 2009; 92: 808–813. 29. Deutch T, Bocca S, Oehninger S, Stadtmauer L, Abuhamad AZ. Magnetic resonance imaging versus three-dimensional transvaginal ultrasound for the diagnosis of ¨ Mullerian anomalies. Fertil Steril 2006; 86: S308. ¨ 30. Olpin JD, Heilbrun M. Imaging of Mullerian duct anomalies. Clin Obstet Gynecol 2009; 52: 40–56. 31. Ludwin A, Ludwin I, Pitynski K, Jach R, Banas T. Are the ESHRE/ESGE criteria of female genital anomalies for diagnosis of septate uterus appropriate? Hum Reprod 2014; 29: 867–868. 32. Grimbizis GF, Gordts S, Di Spiezio Sardo A, Brucker S, De Angelis C, Gergolet M, ¨ Li TC, Tanos V, Brolmann H, Gianaroli L, Campo R. Reply: are the ESHRE/ESGE criteria of female genital anomalies for diagnosis of septate uterus appropriate? Hum Reprod 2014; 29: 868–869.

Ultrasound Obstet Gynecol 2015; 46: 616–622.